Article
Computer Science, Information Systems
Liqun Su, Vincent K. N. Lau
Summary: The study proposes a hierarchical federated learning framework for hybrid data partitioning with a multitier-partitioned neural network architecture. By using primal-dual transform and stochastic coordinate gradient descent ascent algorithm, the training problem is decomposed and updated effectively. Incorporating over-the-air aggregation helps reduce communication complexity and preserve privacy.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Chemistry, Analytical
Miroslav Rujzl, Ladislav Polak, Jiri Petrzela
Summary: This paper presents the design process of a universal development kit based on the concept of an analog computer, capable of modeling complex dynamical systems up to the fourth order. The developed kit includes digital blocks and ADCs/DACs to synthesize mathematical models using segmented piecewise-linear input-output characteristics. The versatility of the kit is demonstrated through experiments ranging from simple oscillators to generators of robust chaotic attractors. Detailed descriptions and technical documentation are provided for the construction of electronic systems and replication of the proposed functional example.
Article
Geosciences, Multidisciplinary
Edoardo Rundeddu, Jose J. Lizarraga, Giuseppe Buscarnera
Summary: This research proposes a hybrid stochastic-mechanical approach to quantify the influence of hydro-mechanical factors on slope stability and rank their importance. Using data from the Campania region in Italy, synthetic intensity-duration thresholds are computed through Monte Carlo simulations. The variability of key variables is evaluated to examine their impact on the results.
Article
Computer Science, Interdisciplinary Applications
Mickael Binois, Robert B. Gramacy
Summary: The study introduces the hetGP package for advanced Gaussian process modeling with input-dependent noise, addressing the complex noise structure and simultaneous modeling of mean and variance fields. The data acquisition is approached using replication and exploration for achieving various goals such as obtaining a globally accurate model, optimization, or contour finding.
JOURNAL OF STATISTICAL SOFTWARE
(2021)
Article
Engineering, Electrical & Electronic
Jacob A. Mueller, Jonathan W. Kimball
Summary: This article proposes a method for predicting the impact of random load behavior on the dynamics of DC microgrids and distribution systems by combining stochastic load models with deterministic microgrid models. The resulting model allows for straightforward calculation of dynamic state moments and assessment of the probability of desirable operating conditions. Specific consideration is given to systems based on the dual active bridge (DAB) topology, with bounds derived for the probability of zero voltage switching (ZVS) in DAB converters. A simple example demonstrates how these bounds can be used to optimize ZVS performance, with predictions verified through Monte Carlo simulations.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2021)
Review
Cell Biology
Scott Alexander Lujan, Thomas A. Kunkel
Summary: This study describes the contribution of DNA mismatch repair (MMR) to the stability of the eukaryotic nuclear genome determined by whole-genome sequencing. While mutation rates in wild-type nuclear genomes are known for over 40 eukaryotic species, fewer measurements have been done in mismatch repair-defective organisms, mainly focusing on Saccharomyces cerevisiae and human tumors. Well-studied organisms include Drosophila melanogaster and Mus musculus, while less genetically tractable species include great apes and long-lived trees. Various techniques have been developed to gather mutation rates per generation or per cell division, and mechanistic studies involving genetic manipulation of MMR genes have been limited to specific organisms like yeast, Arabidopsis thaliana, Caenorhabditis elegans, and one chicken cell line. The few MMR studies reported to date indicate a significant contribution of MMR to genome stability, providing insights that would have been impossible to obtain through reporter gene assays.
Article
Computer Science, Information Systems
Qiuyue Hu, Chaoqin Gan, Xuefeng Liu, Guiyu Gong, Yifan Zhu
Summary: In this paper, a dynamic handover cost model is proposed to address the cost evaluation problem in hybrid VLC/RF heterogeneous networks. The model provides a unified standard for evaluating the performance of handover schemes by evaluating the cost of handover scenarios with different network deployment and user mobility characteristics. The impact of user mobility and network parameters on handover performance is quantified using stochastic geometry and the expressions for the user's cell dwell time and handover rate are derived. Through simulation and analysis, the proposed handover cost model demonstrates better dynamic adaptability and reliability compared to the original model.
Article
Biochemistry & Molecular Biology
Kaixuan Luo, Jianling Zhong, Alexias Safi, Linda K. Hong, Alok K. Tewari, Lingyun Song, Timothy E. Reddy, Li Ma, Gregory E. Crawford, Alexander J. Hartemink
Summary: In this study, we developed a method called TF Occupancy Profiler (TOP) that allows comprehensive observation and quantification of the binding of multiple transcription factors (TFs) in the genome using data from chromatin accessibility experiments. TOP enables the study of how TF occupancy changes across different cell types, exposure to hormones, and genetic backgrounds of individuals.
Article
Engineering, Multidisciplinary
Ch Kuttler, A. Maslovskaya
Summary: Bacterial communication is a complex process that can be formalized and explored through mathematical modeling and computer simulation. A new mathematical model and numerical algorithm have been developed to better describe and study bacterial cooperative behavior.
APPLIED MATHEMATICAL MODELLING
(2021)
Article
Automation & Control Systems
Ying Zhang, Laiqiang Kong
Summary: This paper proposes a hybrid modeling method that combines LSTM and SDE for the prediction of PV output power. Experimental results show the effectiveness of the method and its ability to provide better decision support.
Article
Computer Science, Artificial Intelligence
Yiping Huang, Libao Deng, Jianlei Wang, Weiwei Qiu, Jinfeng Liu
Summary: This study considers uncertainty factors and applies two-stage stochastic programming to model and solve a hybrid flow-shop scheduling problem. A two-stage scenario tree is used to describe the uncertain factors, and a mixed-integer linear programming model is formulated using stochastic programming theory. A novel variant algorithm (H-PDDE) is proposed, which effectively solves the problem by adopting a permutation scheduling decoding method and introducing three new improvement strategies. Computational experiments validate the proposed model and algorithm, showing that the H-PDDE outperforms existing algorithms and conventional PDDE variants in solving the problem more effectively. The stochastic programming model is increasingly superior to the deterministic model in an uncertain environment. The source code and data files of the H-PDDE algorithm can be found at https://github.com/huangyiping-ai/H-PDDE-algorithm.git.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Plant Sciences
Travis Wrightsman, Alexandre P. Marand, Peter A. Crisp, Nathan M. Springer, Edward S. Buckler
Summary: Accessible chromatin regions are critical for gene regulation, and predicting them from sequence in plants has been challenging. By training a deep learning model, we successfully predicted chromatin accessibility and methylation levels across multiple angiosperm species, revealing the conservation of chromatin mechanisms. We also identified important transcription factor families.
Article
Chemistry, Physical
Mengbiao Guo, Zhou Songyang, Yuanyan Xiong
Summary: ChArmTelo is a computational toolbox for analyzing telomeres at the chromosome arm level in humans and other animals. It has been demonstrated to reveal population-specific differences in arm-level telomere lengths and identify liver cancer signatures. It also suggests that disruption of DNA replication origin may play a role in cancer by affecting telomere lengths.
Article
Forestry
Bin Guo, Tingting Chen, Ying Li, Shanwen Li, Wasif Ullah Khan, Ren-Gang Zhang, Kai-Hua Jia, Xinmin An
Summary: Chinese white poplars are dominant in natural ecosystems and widely planted for various purposes. This study analyzed the chloroplast genome and gene expression of an elite individual, identifying the closest related species and detecting key genes involved in translation and photosynthesis. The findings contribute to the understanding of the evolution and molecular genetic improvement of white poplars in China.
Article
Genetics & Heredity
Xun-ju Liu, Xiao-rong Wang, Hao-ru Tang, Qing Chen
Summary: This study reports the complete chloroplast genome sequence of a hybrid blackberry cultivar 'Arapohol', revealing its complex origins. The genome size is 156,621 bp with 134 genes, and phylogenetic analysis shows its close relation to Rubus niveus. The chloroplast genome of this Rubus sp. provides valuable information for understanding the origination of this crop species.
MITOCHONDRIAL DNA PART B-RESOURCES
(2021)